Individual movement patterns in real‐world navigation as a digital marker of cognitive health in the elderly

Abstract

BackgroundDuring normal aging and, particularly, during pathological aging, people often show problems in spatial navigation (SN), rendering it a promising marker of cognitive health in the elderly. However, SN is rarely investigated in people experiencing subjective cognitive decline (SCD), which is considered a very early stage of Alzheimer’s disease. Impaired performance of SCD patients was observed in a computerized navigation task. However, it remains unclear, how this translates to their everyday navigation behavior. Capitalizing on recent advancements in digital technologies, we measured the performance of healthy younger (HYA), healthy older adults (HOA), and SCD patients, when they solve a smartphone‐based navigation task in the real‐world.MethodWe developed a novel smartphone‐based application “Explore” to track GPS data in 67 individuals (HYA, n = 24, HOA, n = 25, SCD, n = 18, still in progress), while they were asked to find five points of interest on the medical campus in Magdeburg. Dynamic time warping was applied to investigate similarities between the raw trajectories. Also, aggregated measures of navigation performance were extracted (e.g., wayfinding distance, wayfinding duration, orientation stops, help usage). Mixed effect models were used to evaluate the effect of group membership on these summary measures. In addition, within‐group heterogeneity in navigation performance was investigated using clustering algorithms.ResultMixed effect models (Figure 1) show that HYA outperform HOA and SCD on all aggregated measures (all p ≤ .018). Between HOA and SCD, no significant differences were found for wayfinding distance and duration (p ≥ .102). However, a trend was detected for amount of help usage (p = .071) and number of orientation stops (p = .057), indicating that SCDs perform slightly worse compared to HOA. The cluster analysis of the movement trajectories revealed that our sample is best described by four navigation patterns (Figure 2).ConclusionWe provide first evidence that movement trajectories from a real‐world navigation task can be clustered into different groups, that could be indicative of age and cognitive health in the elderly. Our preliminary data indicate that certain measures of real‐world navigation behavior might differ between SCD and HOA, which needs to be confirmed in the full sample when data acquisition is complete.